CN109312714B

CN109312714B - Control of a wind turbine taking noise into account

Info

Publication number: CN109312714B
Application number: CN201780034800.7A
Authority: CN
Inventors: K·哈默鲁姆; T·G·霍夫高; E·斯洛斯
Original assignee: Vestas Wind Systems AS
Current assignee: Vestas Wind Systems AS
Priority date: 2016-04-07
Filing date: 2017-04-03
Publication date: 2021-05-04
Anticipated expiration: 2037-04-03
Also published as: CN109312714A; WO2017174089A1; ES2875766T3; US10697431B2; US20190154000A1; EP3440346B1; EP3440346A1

Abstract

The invention relates to a control of a wind turbine taking into account noise measures. Control of a wind turbine is described, wherein a control trajectory is calculated based on a noise measure, which is determined from a predicted operational trajectory. In an embodiment, the predicted operational trajectory is calculated by using a Model Predictive Control (MPC) routine.

Description

Control of a wind turbine taking noise into account

Technical Field

The invention relates to a control of a wind turbine taking into account noise measures.

Background

Typically, the operation of a wind turbine or wind turbine park is aimed at obtaining the maximum gain in capital invested therein, and therefore the wind turbine control system is configured to maximize the output power, i.e. to operate the wind turbine to obtain the maximum power available in the wind, while keeping the wind turbine within operational limits with due regard.

Wind turbines generate noise when operating. This can be problematic for the perimeter of the wind turbine. There are two main sources of noise emitted from wind turbines: noise generated by machine parts of the wind turbine, such as the gearbox and generator, and aerodynamic sounds generated by the air flowing over the rotor blades.

In general, noise generation is dependent on the specific operating parameter settings. However, atmospheric conditions are important for noise propagation in the surrounding environment. Also, factors such as time of day may be important to how sensitive the surrounding environment is to noise.

To address noise emissions, turbines are operated in many countries throughout the world to comply with noise regulations. Such noise regulations may vary from country to country.

Disclosure of Invention

It would be advantageous to control the wind turbine in a manner that takes into account the predicted noise measure during operation of the wind turbine, and which uses the actual operating conditions in conjunction with the continued operation of the wind turbine to determine the predicted noise measure.

Accordingly, in a first aspect, there is provided a method of controlling a wind turbine, comprising:

receiving a current operating state of the wind turbine;

calculating one or more predicted operational trajectories based on the current operational state, the one or more predicted operational trajectories including a predicted control trajectory, wherein the trajectory includes a time series with respect to at least one variable;

determining at least one noise metric based on the at least one predicted operational trajectory;

determining a control trajectory based on at least one noise metric; and

the wind turbine is controlled based on the control trajectory.

In the present invention, the operation of the turbine is based on the calculated control trajectory. A trajectory is a time series of variables during a given time slot that includes the next variable value for the operating parameter associated with the variable and a predicted or expected number of future variable values for the given parameter. For example, the control trajectory may be a pitch trajectory comprising the next pitch command and an expected or predicted number of future pitch commands.

The wind turbine includes a control system for controlling various components of the wind turbine, such as controlling blade pitch settings, power converter setpoints, yaw motors, and the like. The level of noise generated depends on the control action applied.

In the present invention, the control system is arranged for determining at least one noise measure from at least one predicted operational trajectory, i.e. the control system is arranged to determine a predicted noise measure based on actual operational conditions and predicted operation in a future time slot and to control the wind turbine based on such predicted or expected noise measure.

This is advantageous since the expected noise impact on the surroundings can be taken into account directly in the control of the wind turbine during actual operation and on the basis of actual conditions. The noise impact may be taken into account not only for the current operational setting of the wind turbine, but also for the entire duration of the predicted operational trajectory.

In important embodiments, the predicted trajectory is calculated by optimizing at least one cost function, and at least one noise metric is included in the cost function and/or as one or more constraints for the optimization.

Compliance with noise regulations is typically a trade-off between how close to the noise limits of turbine operation and how much the power output is reduced. Since in most cases any measures taken to reduce noise generation will reduce the power output since the turbine needs to operate less aggressively (less aggressive).

Since mathematics deals with tradeoffs in the optimization process through appropriately constructed cost functions and/or constraints, such tradeoffs can be advantageously handled by including noise measures in the cost functions or using noise measures as appropriately set constraints.

In important embodiments, the noise metric may be advantageously calculated in the prediction horizon by calculating one or more predicted operational trajectories using a rolling horizon control routine, such as a Model Predictive Control (MPC) routine.

Other embodiments are described in conjunction with the detailed description.

In a second aspect, the invention also relates to a method of controlling a wind power plant comprising a plurality of wind turbines, the method comprising:

selecting at least one wind turbine of a plurality of wind turbines;

receiving a current operating state of the selected wind turbine;

calculating one or more predicted operational trajectories for the selected wind turbine based on the current operational state, the one or more predicted operational trajectories including a predicted control trajectory, wherein the trajectory includes a time series with respect to at least one variable;

determining at least one noise measure from at least one predicted operational trajectory of the selected wind turbine;

determining a control trajectory based on at least one noise metric; and

the wind turbine is controlled based on the control trajectory.

In this respect, embodiments of the present invention relate to noise handling in wind power plants.

In a further aspect, a control system for wind turbines is provided, comprising a controller for performing the method of the first aspect, and a wind power plant controller arranged for controlling at least selected wind turbines according to the method of the second aspect.

In general, a controller may be a unit or collection of functional units that includes one or more processors, input/output interface(s), and memory capable of storing instructions executable by the processors.

Aspects of the invention may be realized by a computer program product comprising software code adapted for controlling a wind turbine when executed on a data processing system, a controller adapted for a wind turbine, a wind turbine park controller adapted for a wind turbine park controller, the wind turbine park controller being realized to control at least selected turbines of a wind turbine park.

Furthermore, the invention relates to a wind turbine or a collection of wind turbines controlled according to any of the various aspects of the invention.

In general, the various embodiments and aspects of the invention may be combined and coupled in any way possible within the scope of the invention. These and other aspects, features and/or advantages of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter.

Drawings

Embodiments of the invention will now be described, by way of example only, with reference to the accompanying drawings, in which:

FIG. 1 shows a schematic example of a wind turbine;

FIG. 2 shows an exemplary embodiment of a control system and elements of a wind turbine;

FIG. 3 illustrates general aspects of an MPC routine in relation to a running trajectory;

fig. 4 schematically shows the relationship between the generated turbine acoustic power and the acoustic pressure at a given point of acceptance (acceptance point);

FIG. 5 illustrates an example diagram of determining a noise metric from a rolling time domain trajectory;

FIG. 6 schematically shows a wind power plant; and

fig. 7 shows an example of the operation of a wind power plant according to an embodiment.

Detailed Description

Fig. 1 shows an example of a wind turbine 1 in a schematic perspective view. The wind turbine 1 comprises a tower 2, a nacelle 3 arranged at the top end of the tower, and a rotor 4, the rotor 4 being operatively coupled to a generator housed within the nacelle 3. In addition to the generator, the nacelle houses the various components required for converting wind energy into electrical energy, as well as the various components required for operating, controlling and optimizing the performance of the wind turbine 1. The rotor 4 of the wind turbine comprises a central hub 5 and a plurality of blades 6 protruding outwardly from the central hub 5. In the embodiment shown, the rotor 4 comprises three blades 6, but the number may vary. Furthermore, the wind turbine comprises a control system. The control system may be located within the nacelle or distributed at multiple locations within the turbine and communicatively coupled.

FIG. 2 schematically shows an embodiment of the control system 20 and exemplary elements of the wind turbine. The wind turbine comprises rotor blades 6, the rotor blades 6 being mechanically connected to a generator 22 via a gearbox 23. The electric power generated by the generator 22 is input to the grid 24 via the power converter 25. The generator 22 and converter 25 may be based on a full-scale converter (FSC) architecture or a doubly-fed induction generator (DFIG) architecture, although other types may also be used.

The control system includes a number of elements including at least one master controller 200 with a processor and memory such that the processor can perform computational tasks based on instructions stored in the memory. Typically, a wind turbine controller ensures that the wind turbine produces a desired power output level while in operation. This is obtained by adjusting the pitch angle and/or the power extraction of the converter. For this purpose, the control system comprises a pitch system comprising a pitch controller 27 using a pitch reference 28, and a power system comprising a power controller 29 using a power reference 26. The wind turbine rotor comprises rotor blades that can be pitched by a pitch mechanism. The rotor may comprise a collective pitch system, which simultaneously adjusts all pitch angles on all rotor blades, and, in addition, an individual pitch system, which enables individual pitch of the rotor blades.

In an embodiment of the invention, the main controller 200 is programmed to receive the current operating state of the wind turbine. One or more predicted operational trajectories are calculated based on the current operational state, and at least one noise metric is determined based on the at least one predicted operational trajectory. Determining a control trajectory based on the noise metric; and controlling the wind turbine based on the determined control trajectory.

In an embodiment, a predicted operational trajectory and a predicted control trajectory are calculated using a Model Predictive Control (MPC) routine.

FIG. 3 illustrates general aspects of an MPC routine in relation to measured operating variables y and MPC calculated control variables u. The upper part 30 of the diagram shows the state trajectory of the variable y and the lower part 31 of the diagram shows the control trajectory of the controlled variable u.

The trajectory of travel and control may include, but is not limited to, one or more of the following parameters: a pitch value comprising a collective pitch value and an individual pitch value; a rotor speed; rotor acceleration; moving the tower; a power-related parameter; torque related parameters and derivatives of these parameters.

In one embodiment, the operating trajectory is a predicted operating state trajectory. A state is a collection of operating parameters, usually represented as a vector. One exemplary wind turbine state is:

which includes the pitch value theta, the rotor angular velocity omega and the tower top position s, as well as the time derivatives of these parameters. Other and more parameters may be used to define the wind turbine state x. Typically, the running trajectory comprises the running parameters for calculating the desired fatigue load measure.

The state value of the current operational state of the wind turbine may be based on measured sensor readings from sensors arranged to measure sensor data related to the physical state value of the wind turbine. Further, estimated values or calculated values may also be used. In one embodiment, the state may be determined by a state calculator, for example in the form of a dedicated computing unit, such as an observer or kalman filter, which is responsible for determining the current operating state.

The trajectory may also be denoted as a control trajectory. An exemplary control trajectory may be:

which includes a pitch reference signal and a power reference signal. Other and more parameters may be used to define the wind turbine control signal u₁*。

Fig. 3 shows a trace of the measured variable y for a number of discrete time steps (time steps). The figure shows a current time k, as well as a plurality of past time steps 34 and a plurality of future time steps 35 (also referred to as prediction horizon and control horizon for the state variable y and the control variable u, respectively). The known variable values (i.e., the variable values based on the measured values) are marked with filled circles, while the predicted variable values are marked with open circles. One trajectory may comprise a time series of predicted values (i.e. only open circles). The trace need not include past and known values, but may be required in some implementations. In particular, current values 32 may be included for a trace of the measured variable. The trajectory may span a time series of several seconds, such as 5-10 seconds. However, the trajectory may be longer or shorter, depending on the given implementation.

As an example, the trajectory shows the rotor speed ω given a set point to increase the rotor speed. The trajectory shows the current rotor speed 32 as well as the predicted future rotor speed. Also shown are the maximum and minimum values allowed for the variables shown.

Fig. 3 also shows an example of an overall control trajectory determined by using the MPC algorithm. Fig. 3 shows the relationship between the running state trajectory and the overall control trajectory.

While the current kth value is known as the measured variable 32, the current value 37 of the control trajectory is calculated using an MPC routine.

The figure also shows the maximum and minimum allowed values of the control trajectory value of u.

As an example, the trajectory shows the trajectory of the pitch angle, i.e. u ═ θ. Thus, the set point is given to increase the rotor speed and the pitch angle is thereby decreased. The trajectory shows the next pitch setting 37 and the predicted future pitch setting to meet the new set point setting.

MPC is based on iterative finite time domain optimization. At time T, the current state is sampled and a cost minimization control strategy for the future time domain [ T, T + T ] is calculated. Only the first predicted value of the current sample k is used for the control signal, then the turbine state is sampled again and the calculation is repeated starting from the new current state, resulting in a new control trajectory and a new predicted state trajectory. The prediction horizon remains moving forward and therefore the MPC is a rolling horizon controller.

A Model Predictive Control (MPC) controller is a multivariable control algorithm that uses an optimization cost function J over a rolling prediction horizon to calculate optimal control moves.

The optimization cost function may be given by:

referring to FIG. 3, r_iIs the set point of the ith variable, y_iAnd u_iIs the ith trace variable, an

Is a weight matrix defining the relative importance of the variable, and

is a weight matrix that defines a compensation for changes in this variable.

In the above cost function, the noise metric is included in the cost function as a weighted element by the function ρ n (u, y). The weights p may be used to set the importance of the noise metric function in the optimization process.

In one embodiment, the noise metric used in conjunction with cost function optimization may be defined as the acoustic power L_wA。

ρn(u，y)＝ρL_wA(u, y) ═ ρ (f (θ) + g (ω)) equation 2

Acoustic power L_wAMay be expressed as a function of the pitch angle and rotor speed. The pitch angle can be taken into account by the function f (θ), and the rotor speed can be processed by the function g (ω).

FIG. 4 schematically shows the acoustic power L generated by the wind turbine 1_wAThe sound pressure L in a given receiving point can be expressed by using a transmission loss model or a transmission loss function TRL_pA. A specific transmission loss model is defined in the technical standard. An example of such a standard is the ISO 9613-2 standard.

FIG. 5 shows an exemplary diagram of an embodiment showing the determination of a noise measure, where the resulting acoustic power L_wAAnd sound pressure L_pAIs based on the determined rolling temporal trajectory.

In FIG. 3, the predicted time domain 35 of the operational trajectory is shown for rotor speed and pitch angle. These predicted future operational values may be input to the calculation unit 50, possibly together with other input variables, such as wind speed v, to determine a predicted noise level over the predicted time domain. Here, the predicted resulting acoustic power over the prediction time domain 51 is illustrated. In this way, a noise metric for the prediction time domain is determined.

By using the transmission loss function TRL, a noise measure in the prediction time domain can be represented for a given acceptance point 52.

In one embodiment, at least one noise metric is included in the optimization as one or more constraints. This may be included as an alternative to, or in addition to, including the noise measure in the cost function itself. This may depend on the particular embodiment chosen for the optimization problem.

In one embodiment, the optimization criterion for optimizing the at least one cost function is an optimization criterion that keeps the noise measure at a predetermined level. This can be achieved in embodiments by a cost function, by appropriately set constraints or by a combination of both.

Typically, the optimization problem is formulated in terms of an objective function (cost function) and a number of constraints (e.g., maximum/minimum limits, rate of change limits, etc.). When applying such control schemes to normal operation, the objective function is typically formulated to provide a trade-off between noise level and power production, and where certain operating parameters, such as rotor speed, pitch position and speed, and generator torque are limited by constraints.

In one exemplary embodiment, the optimization problem for normal production is of the form:

u^*(t)＝argmin J₀(S(t)，P(t)，u(t))

subject to the following constraints:

ω_R≤Γ_ωR

-5≤θ_i≤90，i∈{1，2，3}

P_E≤3MW

L_pAless than or equal to noise limit

The function argmin is a standard mathematical operator that represents the argument (argument) of the minimum and in which the cost function J is found in the parameter space spanned by S, P, u and t₀The point at which it reaches its minimum. Parameter Γ_ωRRepresenting rated rotor speed, theta_iRepresenting pitch angle (with derivative), and P_ERefers to the rated power of the turbine. The noise constraint is set in such a way that it is clear that the sound pressure should be below a given noise limit in a given acceptance point. This noise limit can be set to a legal threshold.

Here, the nominal cost function J₀The use of the control signal u (t) provides a trade-off between power (P) and load (S) while constraining the limits on rotor speed, blade pitch angle, blade pitch speed, electrical power and resulting acoustic pressure in the predicted time domain. The control signals typically include the blade pitch angle and the power reference of the converter:

by implementing the MPC routine in the controller to calculate the control trajectory, the optimization problem over N time steps (control and prediction horizon) is solved.

As shown in connection with equation 1, the noise metric may be included in the cost function by using the weight ρ. The weight may be used as a sensitivity measure for the noise generated by the wind turbine. For example, the weight may be related to measured atmospheric conditions such as wind speed, rain, snow or other factors affecting noise propagation between the wind turbine and a given acceptance point. This weight may also be related to the measured background noise level, since the wind turbine noise may be less disturbing if the background noise is high.

The weight may also be related to an estimated noise level determined by the wind turbine noise emission model. In this way, if operating parameters are used that do not generate much noise, the weight may be set low or even zero, so that this is not taken into account in the control of the wind turbine. Alternatively, if the noise level generated is high or close to a given threshold, the weight may be set high to ensure that the wind turbine is operating accordingly.

In an embodiment, the weight may be related to a function of the time of day and/or date. In this way, the turbine may, for example, be set to produce more noise during the day or during working days when the environment may be less sensitive to the noise produced, and less noise during the night or other times when the environment is more sensitive to the noise produced.

As shown in fig. 5, the noise level in the prediction time domain may be set to a threshold 53. Such a threshold value may be used as an optimization criterion to keep the generated sound pressure at or below a predetermined level. In one embodiment, the predetermined level is received as input from an external source. Such an external source may be an external microphone or an external unit of noise level or sensitivity to noise monitored at the point of acceptance. The acceptance points may be set continuously to continuously adapt to real-time noise conditions at one or more of the acceptance points.

The predetermined level may be variably set based on various parameters. For example, the predetermined level may be related to a function of time and/or date of day or to a measured background noise level.

In one embodiment, the noise metric may include one or more tonal components. For example, the noise metric may take into account the particular frequency or amplitude of the noise generated. For example, equation 2 may be formulated to include pitch correlation (dependency). Typically, this may be done by including a functional dependence between the rotor speed and/or the set pitch angle at a given frequency, on the level of noise generated. However, it may also be achieved by weighting the noise measure with a given frequency distribution to reduce the noise contribution in certain frequency intervals and to enhance the noise contribution in other frequency intervals. In one embodiment, the turbine may further comprise a look-up table or the like which is issued to provide a weighting factor based on the measured noise to incorporate or replace the weighting factor ρ.

The wind turbine 1 may be comprised in some other wind turbine belonging to a wind power plant 60 (also referred to as a wind farm or wind park).

Fig. 6 schematically shows a wind power plant 60. In one embodiment of the invention, the wind power plant may be controlled by: selecting at least one of the plurality of wind turbines, and calculating a predicted operational trajectory based on the operational state of the selected wind turbine, and determining at least one noise measure from the predicted operational trajectory of the selected wind turbine, and controlling the wind turbine accordingly. The selected turbine may be any, any group or all of the wind turbines in the wind power plant.

Fig. 6 shows two acceptance points. The noise metric may be determined as an aggregate noise metric, such as an aggregate sound pressure, at a specified acceptance point. In this regard, the acoustic power L may be determined for each wind turbine in a first step_wAAnd in a second step the transmission loss function is used separately between each turbine and the acceptance point to determine the aggregate pressure at the selected acceptance point.

In one embodiment, the one or more predicted operational trajectories for the selected at least one wind turbine are calculated by optimizing at least one cost function, and at least one noise measure is included in the cost function and/or at least one noise measure is included in the optimization as one or more constraints in the manner described for a single turbine. By means of a suitably defined cost function and/or suitably set constraints, the operation of individual turbines in a wind park may be continuously adjusted to ensure a specified noise level in one or more acceptance points.

The optimization objective may be to optimize the output power of the wind power plant under the constraint that the sound pressure should be below a legal threshold for the selected acceptance point. Similar to the case of a single turbine, the acceptance points may be externally supplied in a continuous manner from an external source to continuously accommodate real-time noise conditions at one or more of the acceptance points.

Fig. 7 shows a schematic example of the operation of a wind power plant achieving this optimization goal. The figure shows two time intervals. In fig. 7A, in a first time interval 70 the noise level, which is represented as the sound pressure in a given acceptance point, is at a given threshold value 73, whereas in a second time interval 71 the noise level drops below the threshold value, e.g. due to a change in the wind direction. Fig. 7B shows the resulting possible influence of the wind power plant on the output power P. In a first time interval 70, at least some of the individual turbines are derated to ensure that the wind power plant does not exceed a threshold value 73. This results in the aggregate power level P in the first time period being less than the nominal power level 72. In the second time period, the sound pressure at the acceptance point drops below a specified threshold 73 and the wind turbine can thus be operated at its rated output.

Thus, the figure illustrates the trade-off that is formed between the output power and the result produced by the optimization process.

The map is also applicable to single turbine operation, with appropriate modifications.

In an embodiment, the general aspects of the embodiments of the invention may be implemented in a wind turbine power plant controller 61, the wind turbine power plant controller 61 being arranged for controlling one or more wind turbines of a wind power plant. In such an embodiment, the wind farm controller may control one or more selected wind turbines. The wind turbine power plant controller may be implemented in a distributed manner, wherein some parts of the controller are implemented in the wind turbines and other parts of the controller are implemented in the wind power plant controller.

While the invention has been described in connection with specific embodiments, it should not be construed as being limited in any way to the examples given. The invention may be implemented by any suitable means; and the scope of the invention will be understood from the appended claims. Any reference signs in the claims shall not be construed as limiting the scope.

Claims

1. A method of controlling a wind turbine, the method comprising:

receiving a current operating state of the wind turbine;

calculating one or more predicted operational trajectories based on the current operational state, the one or more predicted operational trajectories including a predicted control trajectory, wherein a predicted operational trajectory includes a time series with respect to at least one variable;

determining a control trajectory based on the at least one noise measure; and

controlling the wind turbine based on the control trajectory,

wherein the one or more predicted operational trajectories are calculated by optimizing at least one cost function, and the cost function includes the at least one noise measure, and/or the optimizing includes the at least one noise measure as one or more constraints.

2. The method of claim 1, wherein the one or more predicted trajectories are rolling time domain trajectories having a prediction time domain, and wherein the at least one noise metric is determined for the prediction time domain.

3. The method of claim 1 or 2, wherein the one or more predicted operational trajectories are calculated by using a model predictive control routine.

4. A method according to claim 2, wherein the noise measure is determined by a wind turbine noise emission model for predicting noise of the wind turbine from the one or more operating parameters over the prediction horizon.

5. A method according to claim 1 or 2, wherein the wind turbine noise emission model is arranged for calculating the noise level at a specified acceptance point.

6. The method of claim 1, wherein the cost function comprises a weighted element comprising the at least one noise metric.

7. The method of claim 6, wherein the weight is related to a measured atmospheric condition or the weight is related to a measured background noise level.

8. A method according to claim 6 or 7, wherein the weights are related to an estimated noise level determined by a wind turbine noise emission model.

9. Method according to claim 6 or 7, characterized in that the weights are related to a function of the time of day and/or a function of the date.

10. A method according to claim 1 or 2, wherein the optimization criterion for optimizing the at least one cost function is an optimization criterion that keeps the noise measure at or below a predetermined level.

11. The method of claim 10, wherein the predetermined level is received as input from an external source.

12. The method according to claim 10, characterized in that the predetermined level is related to a function of time of day and/or a function of date, or the predetermined level is related to a measured background noise level.

13. The method of claim 1 or 2, wherein the noise metric comprises one or more tonal components.

14. A method of controlling a wind power plant comprising a plurality of wind turbines, the method comprising:

selecting at least one wind turbine of the plurality of wind turbines;

receiving a current operating state of the selected wind turbine;

calculating one or more predicted operational trajectories for the selected wind turbine based on the current operational state, the one or more predicted operational trajectories including a predicted control trajectory, wherein the predicted operational trajectories include a time series with respect to at least one variable;

determining a control trajectory based on the at least one noise measure; and

controlling the wind turbine based on the control trajectory,

wherein the one or more predicted operational trajectories for the selected at least one wind turbine are calculated by optimizing at least one cost function, and the cost function comprises the at least one noise measure, and/or the optimization comprises the at least one noise measure as one or more constraints.

15. A method according to claim 14, wherein an aggregate noise measure is determined for a given acceptance point based on a plurality of determined noise measures for the selected wind turbine.

16. A control system for a wind turbine, the control system comprising:

a controller arranged to receive a current operating state of the wind turbine and to calculate one or more predicted operating trajectories based on the current operating state, the one or more predicted operating trajectories including a predicted control trajectory, wherein a predicted operating trajectory comprises a time series with respect to at least one variable;

the controller is further configured to determine at least one noise metric from at least one predicted operational trajectory and determine a control trajectory based on the at least one noise metric, wherein the one or more predicted operational trajectories are calculated by optimizing at least one cost function, and the cost function includes the at least one noise metric, and/or the optimization includes the at least one noise metric as one or more constraints; and

controlling the wind turbine based on the control trajectory.

17. A wind power plant controller arranged for controlling one or more wind turbines of a wind power plant comprising a plurality of wind turbines, the wind power plant controller comprising:

a controller arranged to select at least one of the plurality of wind turbines and to receive a current operating state of the selected wind turbine and to calculate one or more predicted operating trajectories for the selected wind turbine based on the current operating state, the one or more predicted operating trajectories including a predicted control trajectory, wherein a predicted operating trajectory comprises a time series with respect to at least one variable;

wherein the controller is further arranged for determining at least one noise measure from at least one predicted operational trajectory of the selected wind turbine, wherein the one or more predicted operational trajectories of the selected at least one wind turbine are calculated by optimizing at least one cost function, and the cost function comprises the at least one noise measure, and/or the optimization comprises the at least one noise measure as one or more constraints; and

controlling at least the selected wind turbine based on the control trajectory.